An Analysis of Volatility Clustering of Equity Factor Strategies

Excerpt

Introduction

Factor investing is a centrepiece of modern portfolio management and there is a lot of research related to it. For example, Jennifer Bender et al. (2013) identified the following equity factors, which historically provide premium and are backed by academic research: Value, Low Size, Low Volatility, High Yield and Momentum. However, there is much more evidence from essential authors Eugene F. Fama and Kenneth R. French in their three-factor (1993) or five-factor (2014) models.

There are many well-documented ways how to allocate among equity factors, besides others, one from our colleague Matus Padysak (2021), who published “The Active vs Passive: Smart Factors, Market Portfolio or Both?”, where he combined smart equity factors with the market portfolio. As he states, the smart equity factors allocated by blended fast/slow time-series momentum signal enhances the market portfolio, especially during market downturns, when they perform very well. Worth mentioning is also a paper from FTSE Russell (2020), which covers allocation of equity factors over three allocation schemes: (1) Equal Exposure, (2) Risk Exposure, (3) Equal Risk Contribution.

Volatility clustering is a well-known effect in equity markets. In simple meaning, volatility clustering refers to a tendency of large changes in asset prices to follow large changes and small changes in asset prices to follow small changes. This interesting effect can be sometimes uncovered as one of the reasons for the functionality of some selected trading strategies. For example, low-volatility months in stock indexes (like the S&P 500 Index) are usually also months with higher performance. As volatility tends to cluster, a low volatility month in the present can signal a low volatility month with a better performance also in the future. This is the basic rationale why long-only trend-following strategies perform pleasantly in bull markets in stocks.

How can we build on this knowledge? It is natural to think also about the other investment universes. Based on this, we will be testing two hypotheses: (1) firstly, if there is a volatility clustering anomaly present in equity factor strategies; (2) secondly, if there is any performance pattern related to volatility. For our equity factor strategies, we chose:

Investment Factor Strategy goes long 20% of firms with the lowest investments (Investments is the growth of total assets for the fiscal year ending in t-1 divided by total assets at the end of t-1) and goes short 20% of firms with the highest investments. The portfolio is rebalanced yearly, and stocks are weighted equally.

Low-Volatility Factor strategy goes long 20% of firms with the lowest average daily volatility during the previous 36-months and goes short 20% of firms with the highest average daily volatility in the last 36-months. The portfolio is rebalanced monthly, and stocks are weighted equally.

Quality Factor Strategy goes long 20% of firms with the highest ROA (Return on assets) calculated as quarterly earnings (Compustat quarterly item IBQ – income before extraordinary items) divided by one-quarter-lagged assets (item ATQ – total assets) and goes short 20% of firms with the lowest quality. The portfolio is rebalanced monthly, and stocks are weighted equally.

Size Factor Strategy goes long 20% of firms with the smallest market capitalization and goes short 20% of firms with the highest market capitalization. The portfolio is rebalanced yearly, and stocks are weighted equally.

Value Factor Strategy goes long 20% of firms with the lowest P/B ratio and goes short 20% of firms with the highest P/B ratio. The portfolio is rebalanced yearly, and stocks are weighted equally.

The investment universe in all five cases consists of 500 US stocks with the highest liquidity (defined as stocks with the highest average monthly dollar turnover). We are interested in factor strategies that are investable in real life; therefore, we are building our investment universe from stocks that have sufficient liquidity.

We use SPY ETF (SPDR S&P 500 ETF Trust) as our proxy for the broad equity market (where the volatility clustering effect is the most visible).

Visit Quantpedia to read the full article:
https://quantpedia.com/an-analysis-of-volatility-clustering-of-equity-factor-strategies/.

Past performance is not necessarily indicative of future results. Any trading symbols displayed are for illustrative purposes only and are not intended to portray recommendations.